Objective: To provide guidelines to the perinatologist regarding extre
mely premature infants based on the experience of the University of To
ronto Newborn Service (two high-risk perinatal units and one outborn n
eonatal intensive care unit), with a catchment area of 60,000 deliveri
es annually. Methods: The study included all births or admissions in t
he Newborn Service from January 1, 1982 to June 30, 1987 with gestatio
nal age determined by the best obstetric estimate of gestational age,
ranging from 23-26 completed weeks. The obstetric records were reviewe
d and the surviving infants followed prospectively for a minimum of 2
years after delivery. Results: Analysis of the neonatal and 2-year fol
low-up data on 568 infants born between 23-26 weeks' gestation reveale
d a 39% mortality rate, which increased with decreasing gestation. The
highest mortality rates occurred following complicated pregnancies, i
ncluding fetal growth restriction. Intact survival increased with incr
easing gestational age, from 11% at 23 weeks to 50% at 26 weeks. There
was a marked improvement in both mortality and morbidity by 25 comple
ted weeks. Conclusions: The results suggest that an aggressive approac
h before 24 completed weeks' gestation is not warranted. From a total
of 60,000 live births per year, only one child born at 23 weeks' gesta
tion and three at 24 weeks were free of major handicap at 2 years.